Isolation of tumor necrosis factor and discovery of its inflammation-promoting effect Beutler's focus on innate immunity began when he was a postdoctoral associate and later an assistant professor in the lab of
Anthony Cerami at
Rockefeller University (1983–1986). Drawing upon skills he had acquired earlier, he isolated mouse "cachectin" from the conditioned medium of LPS-activated mouse macrophages. Cachectin was hypothesized by Cerami to be a mediator of wasting in chronic disease. Its biological activity, the suppression of lipoprotein lipase synthesis in adipocytes, was thought to contribute to wasting, since lipoprotein lipase cleaves fatty acids from circulating triglycerides, allowing their uptake and re-esterification within fat cells. By sequential fractionation of LPS-activated macrophage medium, measuring cachectin activity at each step, Beutler purified cachectin to homogeneity. Determining its N-terminal sequence, he recognized it as mouse
tumor necrosis factor (TNF), and showed that it had strong TNF activity; moreover that human TNF, isolated by a very different assay, had strong cachectin activity. had to that time been defined only by its ability to kill cancer cells. The discovery of a separate role for TNF as a catabolic switch was of considerable interest. Of still greater importance, Beutler demonstrated that TNF acted as a key mediator of endotoxin-induced shock. This he accomplished by raising an antibody against mouse TNF, which he used to neutralize TNF in living mice challenged with lipopolysaccharide (LPS). This was an early hint that TNF might be causally important in
rheumatoid arthritis (as later shown by Feldmann, Brennan, and Maini). Beutler also demonstrated the existence of TNF receptors on most cell types, Before a sensitive immunoassay for TNF was feasible, Beutler used these receptors in a binding competition assay using radio-iodinated TNF as a tracer, which allowed him to precisely measure TNF in biological fluids.
Invention of TNF inhibitors Beutler was recruited to a faculty position at UT Southwestern Medical Center and the Howard Hughes Medical Institute in 1986. Aware that TNF blockade might have clinical applications, he (along with a graduate student, David Crawford, and a postdoctoral associate, Karsten Peppel) invented and patented recombinant molecules expressly designed to neutralize TNF
in vivo (Patent No. US5447851B1). Fusing the binding portion of TNF receptor proteins to the
heavy chain of an immunoglobulin molecule to force receptor dimerization,
Discovery of the LPS receptor, and the role of TLRs in innate immune sensing From the mid-1980s onward Beutler was interested in the mechanism by which LPS activates mammalian immune cells (chiefly
macrophages, but also promoting the well-known adjuvant effect of LPS, and B cell mitogenesis and antibody production. A single, highly specific LPS receptor was presumed to exist as early as the 1960s, based on the fact that allelic mutations in two separate strains of mice, affecting a discrete genetic locus on chromosome 4 termed
Lps, abolished LPS sensing. Although this receptor had been widely pursued, it remained elusive. Beutler reasoned that in finding the LPS receptor, insight might be gained into the first molecular events that transpire upon an encounter between the host and microbial invaders. Utilizing
positional cloning in an effort that began in 1993 and lasted five years, Beutler, together with several postdoctoral associates including Alexander Poltorak, measured TNF production as a qualitative phenotypic endpoint of the LPS response. Analyzing more than 2,000
meioses, they confined the LPS receptor-encoding gene to a region of the genome encompassing approximately 5.8 million base pairs of
DNA. They also showed that while mouse TLR4 is activated by a tetra-acylated LPS-like molecule (lipid IVa), human TLR4 is not, recapitulating the species specificity for LPS partial structures. The structure of the complex, with and without LPS bound, was solved by Jie-Oh Lee and colleagues in 2009. Jules Hoffmann and colleagues had earlier shown that the
Drosophila Toll protein, originally known for its role in embryogenesis, was essential for the
antimicrobial peptide response to fungal infection. However, no molecule derived from
fungi actually became bound to Toll; rather, a proteolytic cascade led to the activation of an endogenous ligand, the protein
Spätzle. This activated
NF-κB within cells of the fat body, leading to antimicrobial peptide secretion. Aware of this work,
Charles Janeway and
Ruslan Medzhitov overexpressed a modified version of human TLR4 (which they called 'h-Toll') and found it capable of activating the transcription factor NF-κB in mammalian cells. They speculated that TLR4 was a "
pattern recognition receptor." However, they provided no evidence that TLR4 recognized any molecule of microbial origin. If a ligand did exist, it might have been endogenous (as in the fruit fly, where Toll recognizes the endogenous protein Spätzle, or as in the case of the IL-1 receptor, which recognizes the endogenous cytokine
IL-1). Indeed, numerous cell surface receptors, including the
TGFβ receptor,
B cell receptor, and
T cell receptor activate NF-κB. In short, it was not clear what TLR4 recognized, nor what its function was. Separate publications, also based on transfection/overexpression studies, held that TLR2 rather than TLR4 was the LPS receptor. The genetic evidence of Beutler and coworkers correctly identified TLR4 as the specific and non-redundant cell surface receptor for LPS, fully required for virtually all LPS activities. This suggested that other TLRs (of which ten are now known to exist in humans) might also act as sensors of infection in mammals, each detecting other signature molecules made by microbes whether or not they were
pathogens in the classical sense of the term. The other TLRs, like TLR4, do indeed initiate innate immune responses. By promoting inflammatory signaling, TLRs can also mediate pathologic effects including
fever, systemic inflammation, and shock. Sterile inflammatory and autoimmune diseases such as
systemic lupus erythematosus also elicit TLR signaling, and disruption of signaling from the nucleic acid sensing TLRs can favorably modify the disease phenotype.
Random Germline Mutagenesis/Forward Genetics in the mouse After completing the positional cloning of the
Lps locus in 1998, Beutler continued to apply a forward genetic approach to the analysis of immunity in mammals. In this process, germline mutations that alter immune function are created in mice through a random process using the alkylating agent
ENU, detected by their phenotypic effects, and then isolated by positional cloning. This work disclosed numerous essential signaling molecules required for the innate immune response, and helped to delineate the biochemistry of innate immunity. Among the genes detected was
Ticam1, implicated by an ENU-induced phenotype called
Lps2. Humans with mutations in
UNC93B1, the human ortholog of the same gene, were subsequently found to be susceptible to recurrent
Herpes simplex virus (HSV)
encephalitis, in which reactivation of latent virus occurs repeatedly in the
trigeminal ganglion at the base of the
midbrain, leading to cortical neuron death. Yet another protein needed to make the endosomal environment suitable for TLR signaling was SLC15A4, identified based on the phenotype
feeble.
feeble was identified in a screen in which immunostimulatory DNA was administered to mice intravenously with measurement of the systemic
type I interferon response. Failure of this response, which is dependent on TLR9 signaling from
plasmacytoid dendritic cells (pDC) was observed in homozygous mutants, and subsequently, failure of TLR7 (but not TLR3) signaling was observed as well. Because the
feeble mutation suppressed SLE in mice, In all, Beutler and colleagues detected 77 mutations in 36 genes in which ENU-induced mutations created defects of TLR signaling, detected due to faulty TNF and/or interferon responses. These genes encoded all TLRs kept under surveillance in screening, all of the four adapter proteins that signal from TLRs, kinases and other signaling proteins downstream,
chaperones needed to escort TLRs to their destinations, proteins that promote the availability of TLR ligands, proteins involved in vesicle transport, and proteins involved in
transcriptional responses to TLR signaling, or the post-translational processing of TNF and/or type I interferons (the proteins assayed in screening). Beutler and colleagues also used ENU mutagenesis to study the global response to a defined infectious agent. They measured susceptibility to mouse
cytomegalovirus (MCMV) and identified numerous genes that make a life-or-death difference during infection, terming this set of genes the MCMV "resistome". These genes were grouped into "sensing," "signaling," "effector," "homeostatic," and "developmental" categories, some of which were wholly unexpected. In the homeostatic category, for example,
Kir6.1 ATP-sensitive potassium channels in the
smooth muscle of the
coronary arteries serve an essential role in the maintenance of blood flow during MCMV infection, and mutations that damage these channels cause sudden death during infection. Other genetic screens in the Beutler laboratory were used to identify genes that mediate homeostatic adaptations of the
intestinal epithelium following a cytotoxic insult; prevent allergic responses, diabetes, or obesity; support normal hematopoiesis; and enable humoral and cellular immunity. Some of these (beginning ~2015) were identified by a new process called automated meiotic mapping, which enabled greatly accelerated mutation identification compared to traditional genetic mapping (see below). In the course of their work, Beutler and his colleagues also discovered genes required for biological processes such as normal
iron absorption,
hearing, pigmentation, metabolism, and
embryonic development. Many human diseases were ultimately linked to variants in the corresponding human genes after initial identification in the mouse by the Beutler laboratory, or by the laboratories of collaborating investigators.
Invention of Automated Meiotic Mapping Prior to 2013, despite the development of methods for
massively parallel sequencing and their application in finding induced germline mutations, positional cloning remained a slow process, limited by the need to genetically map mutations to chromosomal intervals to ascertain which induced mutation (among the average of approximately 60 changes in coding and splicing function induced per pedigree) was responsible for an observed phenotype. This required expansion of a mutant stock, outcrossing to a mapping strain, backcrossing, and genotypic and phenotypic analysis of F2 offspring. Moreover, when phenotypic screening was performed prior to positional cloning, only large effect size mutations (producing essentially qualitative phenotypes) were recoverable. Beutler invented a means of instantly identifying ENU-induced mutations that cause phenotypes. The process, called automated meiotic mapping (AMM), eliminates the need to breed mutant mice to a mapping strain as required in classical
genetic mapping and flags causative mutations as soon as phenotypic assay data are collected. In a laboratory setting, it accelerates positional cloning approximately 200-fold, and permits ongoing measurement of genome saturation as mutagenesis progresses. Not only qualitative phenotypes, but subtle quantitative phenotypes, are detectable and mapped to individual mutations; hence the sensitivity of forward genetics is dramatically increased. AMM depends on statistical computation to detect associations between mutations in either the homozygous or heterozygous state and deviant phenotypes. As of 2022, more than 260,000 ENU-induced non-synonymous coding or splice site mutations had been assayed for phenotypic effects, and more than 5,800 mutations in approximately 2,500 genes had been declared causative of phenotype(s). For certain screens, such as
flow cytometry performed on the blood of germline mutant mice, more than 55% saturation of the genome has been achieved (i.e., more than 55% of all genes in which mutations will create flow cytometric aberrations in the peripheral blood have been detected, most of them based on assessment of multiple alleles, as of July 2021). and metabolism. AMM has also permitted high speed searches for mutations that suppress or augment disease phenotypes; for example, the development of autoimmune (Type 1) diabetes in mice of the NOD strain. ==Awards and recognition==